Genome sequence analysis suggests coevolution of the DIS, SD, and Psi hairpins in HIV-1 genomes.

HIV-1 RNA dimerization is a critical step in viral life cycle. It is a prerequisite for genome packaging and plays an important role in reverse transcription and recombination. Dimerization is promoted by the DIS (dimerization initiation site) hairpin located in the 5' leader of HIV-1 genome. Despite the high genetic diversity in HIV-1 group M, only five apical loops (AAGCGCGCA, AAGUGCGCA, AAGUGCACA, AGGUGCACA and AGUGCAC) are commonly found in DIS hairpins. We refer to the parent DISes with these apical loops as DISLai, DISTrans, DISF, DISMal, and DISC, respectively. Based on identity or similarity of DIS hairpins to parent DISes, we distributed HIV-1 M genomes into five dimerization groups. Comparison of the primary and secondary structures of DIS, SD and Psi hairpins in about 3000 HIV-1 M genomes showed that the mutation frequencies at particular nucleotide positions of these hairpins differ among the dimerization groups, and DISF may be an origin of other parent DISes. We found that DIS, SD and Psi hairpins have hundreds of variants, only some of them occurring rather frequently. The lower part of DIS hairpin with G x AGG internal loop is highly conserved in both HIV-1 and SIV genomes. We supposed that the G-quadruplex, located 56 nts downstream of the Gag start codon, may participate in switching of HIV-1 leader RNA from BMH (branched multiple hairpins) to LDI (long distance interaction) conformation.

Structural transitions in poly(A), poly(C), poly(U), and poly(G) and their possible biological roles.

The homopolynucleotide (homo-oligonucleotide) tracts function as regulatory elements at various stages of mRNAs life cycle. Numerous cellular proteins specifically bind to these tracts. Among them are the different poly(A)-binding proteins, poly(C)-binding proteins, multifunctional fragile X mental retardation protein which binds specifically both to poly(G) and poly(U) and others. Molecular mechanisms of regulation of gene expression mediated by homopolynucleotide tracts in RNAs are not fully understood and the structural diversity of these tracts can contribute substantially to this regulation. This review summarizes current knowledge on different forms of homoribopolynucleotides, in particular, neutral and acidic forms of poly(A) and poly(C), and also biological relevance of homoribopolynucleotide (homoribo-oligonucleotide) tracts is discussed. Under physiological conditions, the acidic forms of poly(A) and poly(C) can be induced by proton transfer from acidic amino acids of proteins to adenine and cytosine bases. Finally, we present potential mechanisms for the regulation of some biological processes through the formation of intramolecular poly(A) duplexes.

Structural model of the complete poly(A) region of HIV-1 pre-mRNA.

In the HIV-1 retrovirus, identical sequences encompassing the AAUAAA hexamer and the U/GU-rich downstream sequence element (DSE) that compose the core poly(A) site are present at both the 5' and 3' ends of the HIV-1 pre-mRNA. The AAUAAA hexamer is partly occluded by base pairing in the upper part of a semi-stable polyA hairpin. This sets the stage for regulation of HIV-1 polyadenylation, which involves reaction suppression at the 5' end and its stimulation at the 3' end. Efficient utilization of the 3' core poly(A) site is promoted by major and minor upstream sequence elements (USEs) which are uniquely present at the 3' end of the HIV-1 transcript. The structures of the HIV-1 5' and 3' poly(A) sites are defined by overall architecture of complete 5' and 3' untranslated terminal regions (UTRs). To our knowledge, there is still no structural model of a complete 3' UTR of the HIV-1 pre-mRNA and complete 3' poly(A) region including the USEs except the fact that the polyA and transactivation response (TAR) hairpins are present at both ends of the HIV-1 pre-mRNA. In this work, we predicted a secondary structure of the 3' UTR of HIV-1 pre-mRNA based on our observation that the minor USEs are located in a region with a high potential to form G-quadruplex structures. We first present structural models for the major USE, complete 3' poly(A) region, and almost entire 3' UTR of HIV-1 pre-mRNA. Our models are built based on the mfold and UNAFold secondary structure prediction of these regions for about 1500 HIV-1 isolates of different subtypes and recombinant forms. We have demonstrated that these models are valid for most of the HIV-1 isolates studied. The proposed models include the known TAR and polyA hairpins and new structural elements containing the U-rich tract of the major USE and U/GU-rich DSE which are fully exposed and accessible to the polyadenylation machinery, which confirms the functional competence of our models.

Downstream elements of mammalian pre-mRNA polyadenylation signals: primary, secondary and higher-order structures.

Primary, secondary and higher-order structures of downstream elements of mammalian pre-mRNA polyadenylation signals [poly(A) signals] are re viewed. We have carried out a detailed analysis on our database of 244 human pre-mRNA poly(A) signals in order to characterize elements in their downstream regions. We suggest that the downstream region of the mammalian pre-mRNA poly(A) signal consists of various simple elements located at different distances from each other. Thus, the downstream region is not described by any precise consensus. Searching our database, we found that approximately 80% of pre-mRNAs with the AAUAAA or AUUAAA core upstream elements contain simple downstream elements, consisting of U-rich and/or 2GU/U tracts, the former occurring approximately 2-fold more often than the latter. Approximately one-third of the pre-mRNAs analyzed here contain sequences that may form G-quadruplexes. A substantial number of these sequences are located immediately downstream of the poly(A) signal. A possible role of G-rich sequences in the polyadenylation process is discussed. A model of the secondary structure of the SV40 late pre-mRNA poly(A) signal downstream region is presented.

What nuclease cleaves pre-mRNA in the process of polyadenylation?

A transcript-specific cleavage by a large set of proteins is the first stage of eukaryotic pre-mRNA polyadenylation. The main participant of this reaction-endonuclease-has not been discovered until now. However, mammalian CPSF-30 and yeast Yth 1p proteins are known to be homologues to Drosophila Clipper (CLP) protein, which possesses endoribonucleolytic activity. In the N-terminal region, all three proteins contain five copies of the CCCH zinc finger motif associated with nucleolytic activity in the case of CLP. The literature data on these proteins are reviewed here. These data were shown not to contradict the hypothesis that CPSF-30 and its homologues are the actual nucleases that cleave pre-mRNA in the process of polyadenylation.